diff --git a/src/BulletCollision/CollisionDispatch/btConvexPlaneCollisionAlgorithm.h b/src/BulletCollision/CollisionDispatch/btConvexPlaneCollisionAlgorithm.h
index 368ca71dd..f49ac45e7 100644
--- a/src/BulletCollision/CollisionDispatch/btConvexPlaneCollisionAlgorithm.h
+++ b/src/BulletCollision/CollisionDispatch/btConvexPlaneCollisionAlgorithm.h
@@ -60,8 +60,8 @@ public:
 		int m_minimumPointsPerturbationThreshold;
 			
 		CreateFunc() 
-			: m_numPerturbationIterations(3),
-			m_minimumPointsPerturbationThreshold(3)
+			: m_numPerturbationIterations(1),
+			m_minimumPointsPerturbationThreshold(1)
 		{
 		}
 		
diff --git a/src/BulletCollision/CollisionShapes/btConvexHullShape.cpp b/src/BulletCollision/CollisionShapes/btConvexHullShape.cpp
index 2b0673677..7b45fad12 100644
--- a/src/BulletCollision/CollisionShapes/btConvexHullShape.cpp
+++ b/src/BulletCollision/CollisionShapes/btConvexHullShape.cpp
@@ -24,12 +24,13 @@ btConvexHullShape ::btConvexHullShape (const btScalar* points,int numPoints,int
 	m_shapeType = CONVEX_HULL_SHAPE_PROXYTYPE;
 	m_unscaledPoints.resize(numPoints);
 
-	unsigned char* pointsBaseAddress = (unsigned char*)points;
+	unsigned char* pointsAddress = (unsigned char*)points;
 
 	for (int i=0;i<numPoints;i++)
 	{
-		btVector3* point = (btVector3*)(pointsBaseAddress + i*stride);
-		m_unscaledPoints[i] = point[0];
+		btScalar* point = (btScalar*)pointsAddress;
+		m_unscaledPoints[i] = btVector3(point[0], point[1], point[2]);
+		pointsAddress += stride;
 	}
 
 	recalcLocalAabb();
diff --git a/src/BulletCollision/CollisionShapes/btStridingMeshInterface.cpp b/src/BulletCollision/CollisionShapes/btStridingMeshInterface.cpp
index 752a5fc1d..b3243bf44 100644
--- a/src/BulletCollision/CollisionShapes/btStridingMeshInterface.cpp
+++ b/src/BulletCollision/CollisionShapes/btStridingMeshInterface.cpp
@@ -23,66 +23,127 @@ btStridingMeshInterface::~btStridingMeshInterface()
 
 void	btStridingMeshInterface::InternalProcessAllTriangles(btInternalTriangleIndexCallback* callback,const btVector3& aabbMin,const btVector3& aabbMax) const
 {
-	(void)aabbMin;
-	(void)aabbMax;
-	int numtotalphysicsverts = 0;
-	int part,graphicssubparts = getNumSubParts();
-	const unsigned char * vertexbase;
-	const unsigned char * indexbase;
-	int indexstride;
-	PHY_ScalarType type;
-	PHY_ScalarType gfxindextype;
-	int stride,numverts,numtriangles;
-	int gfxindex;
-	btVector3 triangle[3];
-	btScalar* graphicsbase;
+   (void)aabbMin;
+   (void)aabbMax;
+   int numtotalphysicsverts = 0;
+   int part,graphicssubparts = getNumSubParts();
+   const unsigned char * vertexbase;
+   const unsigned char * indexbase;
+   int indexstride;
+   PHY_ScalarType type;
+   PHY_ScalarType gfxindextype;
+   int stride,numverts,numtriangles;
+   int gfxindex;
+   btVector3 triangle[3];
 
-	btVector3 meshScaling = getScaling();
+   btVector3 meshScaling = getScaling();
 
-	///if the number of parts is big, the performance might drop due to the innerloop switch on indextype
-	for (part=0;part<graphicssubparts ;part++)
-	{
-		getLockedReadOnlyVertexIndexBase(&vertexbase,numverts,type,stride,&indexbase,indexstride,numtriangles,gfxindextype,part);
-		numtotalphysicsverts+=numtriangles*3; //upper bound
+   ///if the number of parts is big, the performance might drop due to the innerloop switch on indextype
+   for (part=0;part<graphicssubparts ;part++)
+   {
+      getLockedReadOnlyVertexIndexBase(&vertexbase,numverts,type,stride,&indexbase,indexstride,numtriangles,gfxindextype,part);
+      numtotalphysicsverts+=numtriangles*3; //upper bound
 
-		switch (gfxindextype)
-		{
-		case PHY_INTEGER:
-			{
-				for (gfxindex=0;gfxindex<numtriangles;gfxindex++)
-				{
-					unsigned int* tri_indices= (unsigned int*)(indexbase+gfxindex*indexstride);
-					graphicsbase = (btScalar*)(vertexbase+tri_indices[0]*stride);
-					triangle[0].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),graphicsbase[2]*meshScaling.getZ());
-					graphicsbase = (btScalar*)(vertexbase+tri_indices[1]*stride);
-					triangle[1].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
-					graphicsbase = (btScalar*)(vertexbase+tri_indices[2]*stride);
-					triangle[2].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
-					callback->internalProcessTriangleIndex(triangle,part,gfxindex);
-				}
-				break;
-			}
-		case PHY_SHORT:
-			{
-				for (gfxindex=0;gfxindex<numtriangles;gfxindex++)
-				{
-					unsigned short int* tri_indices= (unsigned short int*)(indexbase+gfxindex*indexstride);
-					graphicsbase = (btScalar*)(vertexbase+tri_indices[0]*stride);
-					triangle[0].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),graphicsbase[2]*meshScaling.getZ());
-					graphicsbase = (btScalar*)(vertexbase+tri_indices[1]*stride);
-					triangle[1].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
-					graphicsbase = (btScalar*)(vertexbase+tri_indices[2]*stride);
-					triangle[2].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
-					callback->internalProcessTriangleIndex(triangle,part,gfxindex);
-				}
-				break;
-			}
-		default:
-			btAssert((gfxindextype == PHY_INTEGER) || (gfxindextype == PHY_SHORT));
-		}
+///unlike that developers want to pass in double-precision meshes in single-precision Bullet build
+///so disable this feature by default
+///see patch http://code.google.com/p/bullet/issues/detail?id=213
 
-		unLockReadOnlyVertexBase(part);
-	}
+#ifdef BT_USE_DOUBLE_PRECISION
+      switch (type)
+      {
+         case PHY_FLOAT:
+#endif
+		 {
+
+            float* graphicsbase;
+            
+            switch (gfxindextype)
+            {
+               case PHY_INTEGER:
+               {
+                  for (gfxindex=0;gfxindex<numtriangles;gfxindex++)
+                  {
+                     unsigned int* tri_indices= (unsigned int*)(indexbase+gfxindex*indexstride);
+                     graphicsbase = (float*)(vertexbase+tri_indices[0]*stride);
+                     triangle[0].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (float*)(vertexbase+tri_indices[1]*stride);
+                     triangle[1].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (float*)(vertexbase+tri_indices[2]*stride);
+                     triangle[2].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
+                     callback->internalProcessTriangleIndex(triangle,part,gfxindex);
+                  }
+                  break;
+               }
+               case PHY_SHORT:
+               {
+                  for (gfxindex=0;gfxindex<numtriangles;gfxindex++)
+                  {
+                     unsigned short int* tri_indices= (unsigned short int*)(indexbase+gfxindex*indexstride);
+                     graphicsbase = (float*)(vertexbase+tri_indices[0]*stride);
+                     triangle[0].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (float*)(vertexbase+tri_indices[1]*stride);
+                     triangle[1].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (float*)(vertexbase+tri_indices[2]*stride);
+                     triangle[2].setValue(graphicsbase[0]*meshScaling.getX(),graphicsbase[1]*meshScaling.getY(),	graphicsbase[2]*meshScaling.getZ());
+                     callback->internalProcessTriangleIndex(triangle,part,gfxindex);
+                  }
+                  break;
+               }
+               default:
+                  btAssert((gfxindextype == PHY_INTEGER) || (gfxindextype == PHY_SHORT));
+            }
+            break;
+         }
+
+#ifdef BT_USE_DOUBLE_PRECISION
+         case PHY_DOUBLE:
+         {
+            double* graphicsbase;
+
+            switch (gfxindextype)
+            {
+               case PHY_INTEGER:
+               {
+                  for (gfxindex=0;gfxindex<numtriangles;gfxindex++)
+                  {
+                     unsigned int* tri_indices= (unsigned int*)(indexbase+gfxindex*indexstride);
+                     graphicsbase = (double*)(vertexbase+tri_indices[0]*stride);
+                     triangle[0].setValue((btScalar)graphicsbase[0]*meshScaling.getX(),(btScalar)graphicsbase[1]*meshScaling.getY(),(btScalar)graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (double*)(vertexbase+tri_indices[1]*stride);
+                     triangle[1].setValue((btScalar)graphicsbase[0]*meshScaling.getX(),(btScalar)graphicsbase[1]*meshScaling.getY(),  (btScalar)graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (double*)(vertexbase+tri_indices[2]*stride);
+                     triangle[2].setValue((btScalar)graphicsbase[0]*meshScaling.getX(),(btScalar)graphicsbase[1]*meshScaling.getY(),  (btScalar)graphicsbase[2]*meshScaling.getZ());
+                     callback->internalProcessTriangleIndex(triangle,part,gfxindex);
+                  }
+                  break;
+               }
+               case PHY_SHORT:
+               {
+                  for (gfxindex=0;gfxindex<numtriangles;gfxindex++)
+                  {
+                     unsigned short int* tri_indices= (unsigned short int*)(indexbase+gfxindex*indexstride);
+                     graphicsbase = (double*)(vertexbase+tri_indices[0]*stride);
+                     triangle[0].setValue((btScalar)graphicsbase[0]*meshScaling.getX(),(btScalar)graphicsbase[1]*meshScaling.getY(),(btScalar)graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (double*)(vertexbase+tri_indices[1]*stride);
+                     triangle[1].setValue((btScalar)graphicsbase[0]*meshScaling.getX(),(btScalar)graphicsbase[1]*meshScaling.getY(),  (btScalar)graphicsbase[2]*meshScaling.getZ());
+                     graphicsbase = (double*)(vertexbase+tri_indices[2]*stride);
+                     triangle[2].setValue((btScalar)graphicsbase[0]*meshScaling.getX(),(btScalar)graphicsbase[1]*meshScaling.getY(),  (btScalar)graphicsbase[2]*meshScaling.getZ());
+                     callback->internalProcessTriangleIndex(triangle,part,gfxindex);
+                  }
+                  break;
+               }
+               default:
+                  btAssert((gfxindextype == PHY_INTEGER) || (gfxindextype == PHY_SHORT));
+            }
+            break;
+         }
+         default:
+            btAssert((type == PHY_FLOAT) || (type == PHY_DOUBLE));
+      }
+#endif //BT_USE_DOUBLE_PRECISION
+
+      unLockReadOnlyVertexBase(part);
+   }
 }
 
 void	btStridingMeshInterface::calculateAabbBruteForce(btVector3& aabbMin,btVector3& aabbMax)
diff --git a/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.cpp b/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.cpp
index 5f2925328..b3527da89 100644
--- a/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.cpp
+++ b/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.cpp
@@ -44,11 +44,9 @@ void	btTriangleIndexVertexArray::getLockedVertexIndexBase(unsigned char **vertex
 
 	numverts = mesh.m_numVertices;
 	(*vertexbase) = (unsigned char *) mesh.m_vertexBase;
-	#ifdef BT_USE_DOUBLE_PRECISION
-	type = PHY_DOUBLE;
-	#else
-	type = PHY_FLOAT;
-	#endif
+
+   type = mesh.m_vertexType;
+
 	vertexStride = mesh.m_vertexStride;
 
 	numfaces = mesh.m_numTriangles;
@@ -64,11 +62,9 @@ void	btTriangleIndexVertexArray::getLockedReadOnlyVertexIndexBase(const unsigned
 
 	numverts = mesh.m_numVertices;
 	(*vertexbase) = (const unsigned char *)mesh.m_vertexBase;
-	#ifdef BT_USE_DOUBLE_PRECISION
-	type = PHY_DOUBLE;
-	#else
-	type = PHY_FLOAT;
-	#endif
+
+   type = mesh.m_vertexType;
+   
 	vertexStride = mesh.m_vertexStride;
 
 	numfaces = mesh.m_numTriangles;
diff --git a/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.h b/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.h
index eb79ff5e4..c61be4cff 100644
--- a/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.h
+++ b/src/BulletCollision/CollisionShapes/btTriangleIndexVertexArray.h
@@ -25,18 +25,33 @@ subject to the following restrictions:
 ///Instead of the number of indices, we pass the number of triangles.
 ATTRIBUTE_ALIGNED16( struct)	btIndexedMesh
 {
-	BT_DECLARE_ALIGNED_ALLOCATOR();
+   BT_DECLARE_ALIGNED_ALLOCATOR();
 
-	int			m_numTriangles;
-	const unsigned char *		m_triangleIndexBase;
-	int			m_triangleIndexStride;
-	int			m_numVertices;
-	const unsigned char *		m_vertexBase;
-	int			m_vertexStride;
-	// The index type is set when adding an indexed mesh to the
-	// btTriangleIndexVertexArray, do not set it manually
-	PHY_ScalarType		m_indexType;
-	int			pad;
+   int                     m_numTriangles;
+   const unsigned char *   m_triangleIndexBase;
+   int                     m_triangleIndexStride;
+   int                     m_numVertices;
+   const unsigned char *   m_vertexBase;
+   int                     m_vertexStride;
+
+   // The index type is set when adding an indexed mesh to the
+   // btTriangleIndexVertexArray, do not set it manually
+   PHY_ScalarType m_indexType;
+
+   // The vertex type has a default type similar to Bullet's precision mode (float or double)
+   // but can be set manually if you for example run Bullet with double precision but have
+   // mesh data in single precision..
+   PHY_ScalarType m_vertexType;
+
+
+   btIndexedMesh()
+      {
+#ifdef BT_USE_DOUBLE_PRECISION
+      m_vertexType = PHY_DOUBLE;
+#else // BT_USE_DOUBLE_PRECISION
+      m_vertexType = PHY_FLOAT;
+#endif // BT_USE_DOUBLE_PRECISION
+      }
 }
 ;
 
@@ -69,10 +84,10 @@ public:
 	//just to be backwards compatible
 	btTriangleIndexVertexArray(int numTriangles,int* triangleIndexBase,int triangleIndexStride,int numVertices,btScalar* vertexBase,int vertexStride);
 	
-	void	addIndexedMesh(const btIndexedMesh& mesh, PHY_ScalarType indexType = PHY_INTEGER)
+   void  addIndexedMesh(const btIndexedMesh& mesh, PHY_ScalarType indexType = PHY_INTEGER)
 	{
-		m_indexedMeshes.push_back(mesh);
-		m_indexedMeshes[m_indexedMeshes.size()-1].m_indexType = indexType;
+      m_indexedMeshes.push_back(mesh);
+      m_indexedMeshes[m_indexedMeshes.size()-1].m_indexType  = indexType;
 	}
 	
 	
diff --git a/src/BulletDynamics/ConstraintSolver/btContactSolverInfo.h b/src/BulletDynamics/ConstraintSolver/btContactSolverInfo.h
index e99430c00..3293c8de9 100644
--- a/src/BulletDynamics/ConstraintSolver/btContactSolverInfo.h
+++ b/src/BulletDynamics/ConstraintSolver/btContactSolverInfo.h
@@ -77,7 +77,7 @@ struct btContactSolverInfo : public btContactSolverInfoData
 		m_splitImpulsePenetrationThreshold = -0.02f;
 		m_linearSlop = btScalar(0.0);
 		m_warmstartingFactor=btScalar(0.85);
-		m_solverMode = SOLVER_USE_WARMSTARTING | SOLVER_SIMD ;//SOLVER_RANDMIZE_ORDER
+		m_solverMode = SOLVER_USE_WARMSTARTING | SOLVER_SIMD;// | SOLVER_RANDMIZE_ORDER;
 		m_restingContactRestitutionThreshold = 2;//resting contact lifetime threshold to disable restitution
 	}
 };
diff --git a/src/BulletDynamics/ConstraintSolver/btSequentialImpulseConstraintSolver.cpp b/src/BulletDynamics/ConstraintSolver/btSequentialImpulseConstraintSolver.cpp
index 685a812d4..373cdfa26 100644
--- a/src/BulletDynamics/ConstraintSolver/btSequentialImpulseConstraintSolver.cpp
+++ b/src/BulletDynamics/ConstraintSolver/btSequentialImpulseConstraintSolver.cpp
@@ -490,7 +490,7 @@ void	btSequentialImpulseConstraintSolver::convertContact(btPersistentManifold* m
 
 
 				///warm starting (or zero if disabled)
-				if (0)//infoGlobal.m_solverMode & SOLVER_USE_WARMSTARTING)
+				if (infoGlobal.m_solverMode & SOLVER_USE_WARMSTARTING)
 				{
 					solverConstraint.m_appliedImpulse = cp.m_appliedImpulse * infoGlobal.m_warmstartingFactor;
 					if (rb0)